We have been interested in the protein CTCF, which we first identified some years ago as having properties of an insulator, blocking interaction between enhancers and promoters when placed between them. We demonstrated that this activity plays an important role in regulating parent of origin allele-specific gene expression at the Igf2/H19 imprinted locus. Work in recent years has shown that a principal mode of action of CTCF is to stabilize interactions between CTCF binding sites on DNA, leading to formation of loop domains. Depending on the geometry of the interactions such loops can either exclude an enhancer leading to insulation, or bring enhancer and promoter closer together, leading to activation. In the course of searching for CTCF interacting proteins we identified a strong interaction, both in vitro and in vivo, between CTCF and the centromeric protein CENP-E. We were able to identify, within the repeat DNA sequences of the pericentromeric region, unusual CTCF sites that recruited CENP-E. Binding of CENP-E is dependent on binding of CTCF. These results suggest a novel function for CTCF in recruiting CENP-E to the centromere, a novel role for these unusual CTCF binding sites, and perhaps a novel and functionally important organized structure centered on CTCF. We have recently reported an interaction between CTCF and the DEAD box helicase, p68. We found that a long non-coding RNA, SRA, which forms a complex with p68, is essential for this interaction. We have investigated the nature of the p68-SRA interaction, and the role of the RNA in insulator activity of CTCF. We have now extended these studies to a detailed examination of SRA function. Many long non-coding RNAs recruit either the Polycomb complex, PRC2, which delivers histone modifications that are associated with silencing of gene expression, or one of the Trithorax group related complexes, which delivers activating modifications. We find that a single SRA molecule can bind both complexes simultaneously, a novel function for long non-coding RNAs. This result suggests that sites on DNA at which SRA is bound could display both silencing and activating marks. Such bivalent sites are known to occur in pluripotent stem cells and we are exploring the possibility that SRA is involved in creation of such sites. All of these results relate to the role of chromatin structure, histone modifications, and long range organization of the genome in cell function, and are in turn related to questions of normal and abnormal cell metabolism and cell division.